st: recursive variable

[ulio Glogowsky <ulioglogowsky@gmail.com>]

Dear Statalist,

I am trying to create two variables which are dependent on its lagged
values and on a lagged value of a second variable:

y = b1*y(t-1)+b2*x(t-1)+u
x = b1*x(t-1)+b2x(t-1)+u

I am familiar with the 'one variable solution' posted by Nick Cox, as
well as with the paper by Roger Newson's useful Stata Tip in Stata
Journal 4(4) 2004:

//tsset group time
//gen x = y
//replace x = constant * L.x

//gen x = y
//bysort group (time): replace x = constant * x[_n-1] if _n > 2

If anyone has an idea on how to do this with more than 1 variable, I
would be most grateful.


In addition I'm trying to generalize the solution to the 'variable lag
case'. For this purpose I would like to use coefficient vectors and
lists:

Illustrative Example
gen local `list' "y(t-1) x(t-1)"
B = (b1, b2)

replace y = B'* `list' +U

Would it be possible to multiply vectors with variable lists (e.g. in Mata?)?

Thanks
Uli
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/